Anomaly detection method and recording medium

Abstract

An anomaly detection method includes generating second data by adding normal noise to first data; generating third data by adding abnormal noise to the first data; inputting the first data, the second data, and the third data to a neural network; calculating a first normal score, a second normal score, and a third normal score; calculating a third difference based on a first difference and a second difference, the first difference being based on a difference between the first normal score and the second normal score, the second difference being based on a difference between the first normal score and the third normal score; changing the neural network so that the third difference becomes smallest; inputting, to the changed neural network, fourth data that is unknown in terms of whether the fourth data is normal or abnormal; and determining whether the fourth data is normal or abnormal.

Description

BACKGROUND1. Technical Field[0001]The present disclosure relates to an anomaly detection method using a neural network and a recording medium.2. Description of the Related Art[0002]Pascal Vincent, “A Connection Between Score Matching and Denoising Autoencoders”, Neural Computation, July 2011, Vol. 23, Issue 7, pp. 1661-1674 (hereinafter referred to as Non-Patent Document 1) discloses a relationship between a learning method called score matching and a learning method called denoising autoencoders. Score matching is advantageous in being applicable to various models and is disadvantageous in having difficulty in learning. However, Non-Patent Document 1 shows that the solution obtained from score matching and the solution obtained from denoising autoencoders match each other. Thus, learning can be easily performed with the advantage of being applicable to various models being maintained, by expressing models using score matching and by performing learning using denoising autoencoders....

AI Technical Summary

Benefits of technology

The patent is about a method and system for detecting abnormal situations using a neural network. The invention aims to improve the accuracy of anomaly detection. Its technical effects include improving the accuracy of detecting abnormal situations using a neural network, which can help in identifying potential threats and prevent damage or injury.

Problems solved by technology

Score matching is advantageous in being applicable to various models and is disadvantageous in having difficulty in learning.

Application of the techniques described in Non-Patent Documents 1 and 2 to anomaly detection using the neural network described in Patent Document 1 may cause a decrease in accuracy.

Images